← All projects

NBA Shot Efficiency Pareto Model

Completed

Treats the basketball court as a portfolio to find the optimal risk/reward frontier for shot selection.

Pareto / Multi-Objective OptimizationStatistical AnalysisPythonData Analysis

A mathematical optimization project inspired by the "Moreyball" analytics revolution, which frames shot selection as a multi-objective problem: maximizing expected value while accounting for variance. By treating the court as an asset portfolio, the model calculates the exact break-even frontier where a team cannot increase expected point return without increasing the risk of missing the shot — directly applying Pareto optimization to sports analytics. Zone-level efficiency is trained on 218,701 logged shots from the 2023–24 NBA season; a separate era-comparison view plots the 100 highest-scoring player-seasons since 1979–80 — the first season with a 3-point line, and 47 seasons of box-score data — to show how the risk/reward frontier has drifted as the league has shifted toward 3-point shooting.